Earlier attemts at project scanners are flawed - they simply create a database of loose resources and perform a simple text search on every byte in each file. These databases are poorly organized, meaning scans take longer than they should for what should be a simple scan. In addition, since these are simple text scans they will also include text such as hero names and dialogue in events as file names, inflating the detected resources list with unneccessary files. Older scanners only scan a single directory for resources - and usually this is hardcoded to the installed RTP directory. The final nail in the coffin is that while these scanners can detect if a named resource is in your project directory and exclude it from an import list, it does not say whether or not those resources are actually used in the project in question.

RM-Tool changes all four of these complaints and adds in extra functions that make it easier to use.

While it still has to examine every file in the project, it only needs to scan a given file once to determine the files that are used. Each file is organized by type, meaning the file check proceeds only against files of the same type. Files only need to be checked once and are only checked at the end of the scan - a decently large project can be scanned a matter of seconds, while older scanners would take 10-15 minutes to complete a job of the same complexity. Only the neccessary blocks in each file are scanned - the rest are skipped and ignored. Dialogue and the like has no impact on the scanner, and this helps the speed as well. Files can be imported from multiple directories from anywhere on your computer and these directories are imported from in order. Finally, it also detects if you're not using specific files in your project and can automatically delete them.